irm

© All Rights Reserved

Просмотров: 7

irm

© All Rights Reserved

- Ultrasound 120925142801 Phpapp02
- Alltexts.2011
- Mri Syllabus
- EFW-2301 EFilm Workstation 4.2.x Release Notes
- Radiologi Stroke
- Anticancer Drug Testing- Proteome Signatures in Breast by MALDI, Immunostaining MRI-PET Biomarker
- msw_I
- Lecture18 Magnetic Magnetic Field of a Current Loop; Ampere's Law; Solenoids; Toroids
- Primary Hyperparathyroidism 1
- Brain Abscess Seminar 2000
- dwillis form63
- Orden 2012 NeuroImage
- Cristina 3
- Kha Chan
- WJR-6-794
- 1983-10_HP Journal Papers
- Master Biomedical Engineering Guide 2015-2016 - Version 20151209
- Radiology Procedure Manual-sampls
- MAGNET 2 .docx
- Lecture 1 Introduction Digital Image Processing

Вы находитесь на странице: 1из 12

6, 2334, 2009

Z. Ren, D. Xie, and H. Li

Shenyang University of Technology

Shenyang 110023, China

AbstractThe shimming method used for producing high field

homogeneity of the open permanent main magnet for magnetic

resonance imaging (MRI) is researched in this paper. The central

shimming method based on integer programming is proposed, which

fulfills the combination of optimal theory and the practical manual

shimming. The formulation of shimming is solved by using Lingo

software and the numerical analysis method is used to compute the

contribution of small shim arrays. The homogeneity of imaging

region is eventually advanced by 50%. The validity of the method

is verified by using simulation test of two magnets with different basic

magnetic fields. The efficiency of shimming is improved through actual

measuring experiment corporated with the manufacturing enterprise.

1. INTRODUCTION

Owing to its merits of high resolution and high imaging definition,

especially for display of soft tissue, the technology of magnetic

resonance imaging (MRI) has developed rapidly since 1980s, and the

MRI device has become one of the five indispensable clinical medical

imaging equipments now. The others include the ultrasonic, radiology,

X-ray and nuclear instruments. In order to obtain top-quality images,

the high homogeneity of static magnetic field in the imaging region is

required. A magnet system called main magnet is carefully designed to

produce the specified magnetic field. However, for the open permanent

main magnet, the field homogeneity is far from the imaging criterion

due to the influence of the manufacturing tolerances, temperature,

and surrounding objects after magnet is assembled. Therefore, certain

Corresponding author: Z. Ren (rzyhenan@163.com).

24

magnetic field uniformities to an acceptable level for imaging.

There are two shimming methods, active shimming and passive

shimming. The former employs the magnetic field created by currentcarrying coils to offset the inhomogeneous magnetic field components

of special harmonics in the imaging region [1, 2]. As for the latter,

many small shims made of permanent magnet are fixed on the two

pole-surfaces of the main magnet [3]. The inhomogeneous magnetic

component in the imaging region can be compensated through

the magnetic field created by the small magnet sheet array [4, 5].

The shimming technique for open permanent MRI main magnet is

researched and passive shimming is used in this paper.

Integer programming (IP) is firstly introduced into shimming of

MRI magnet by B. Dorri [5], which establishes the groundwork for

succeeding research. A mathematic model of passive shimming based

on linear integer programming method was proposed by reference [6], in

which the computation results for the positions and numbers of shims

were shown by figures as a clear instruction to the manual shimming

process. The method described in [6] uses less shims, less time and

is with less dependence on the operators experience. However, the

negative shim, which means the shim with its magnetic polarity

opposite to that of the main magnet, is not considered in the method,

which causes that the homogeneity of the main magnetic field is hardly

improved after obtaining an amelioration of 20% from the initial status.

A new central shimming method based on the study of the open

permanent magnet system for MRI is proposed in this paper. The

combination of integer programming algorithm and manual shimming

mode based on the actual measurement is fulfilled. The positions and

numbers of shims are optimized based on the magnetic flux density

on the longitudinal symmetrical axis in the imaging region. Therefore

the inhomogeneous magnetic field components can be offset, and the

difference of magnetic flux densities on the sampling plane and the

symmetrical axis is reduced. The formulation and implementation of

the method proposed are described in detail in the following sections,

and the validity is verified by simulation test of two magnets, one is

called magnet 1 whose height between the up and down magnet pole

is 500 mm, the diameter of pole face is 800 mm, the other is magnet 2

with the height of 550 mm and diameter of 1000 mm. The samariumcobalt magnetic shims used in the shimming process include 5 types.

The diameters of shims are 15 mm, 12 mm, 9 mm, 7 mm and 5 mm

separately. The heights are 1.5 mm, 1.5 mm, 1 mm, 1 mm and 1 mm

separately.

25

2. PROCESS OF PASSIVE SHIMMING

The principal structure of an open permanent main magnet for MRI

is shown in Fig. 1. The imaging region is a spherical volume with its

diameter of 400 mm. The inhomogeneity, , of the magnetic field in

the region is defined as

=

Bmax Bmin

106

Bmean

(1)

Bmax and Bmin are the maximum and minimum of the magnetic flux

density in the region respectively, while Bmean is the average value of

the magnetic flux density in the imaging region.

With the manual shimming, the position and number of

compensating shims are tentatively determined by shimming engineer.

However, the process is completed by at least two engineers spending

56 days and using thousands of shims. Besides, there are many shims

used, and their distribution is irregular shown in Fig. 2. The amount

of work is heavy and low efficiency in the process of manual shimming,

hence its improvement is of significance.

3. FORMULATION OF THE METHOD PROPOSED

There is no current supply in the imaging region, so that the scalar

magnetic potential satisfies the Laplace equation

2 = 0

(2)

26

that is, is a harmonic function. According to the properties of the

harmonic function, the maximum and minimum of the function are

all located on the boundary of the region. Thus the practical target

area on which the homogeneity of magnetic field has to be checked is

reduced and restricted onto the boundary of the region, the surface of

the sphere.

After the main magnet is assembled, the magnetic field on the

spherical surface of the imaging region is measured firstly. The

sampling values of the magnetic field are considered as basic field,

which are measured using the magnetic field density meter PT2025,

whose measurement range is from 0.043 T to 13.7 T, and its precision

is 107 T. The locations of the sampling points are shown in Fig. 3.

The whole spherical surface is divided into 13 layers, and 12 sampling

points located at equal interval are set on each layer, so that the

data of 156 sampling points are gained. The passive shimming is

programmed based on these data. In Fig. 3, which is the front view of

the imaging sphere. Bm is the magnetic flux density on the longitudinal

symmetrical axis and defined as the central magnetic field in the

following text; Bi is the basic field value of the ith point on the sphere

surface.

During the practical manual shimming process, Bm is compared

with Bi . On the same layer, if Bi is less than Bm , positive shim

whose magnetic polarity is opposite to the negative shim is set on the

corresponding pole surface so that the magnetic field around point i can

be enhanced; whereas, the negative shim is set on the corresponding

pole surface so that the magnetic field around point i can be reduced.

The equatorial plane of the sphere is selected as an interface, the

27

upper hemispherical shimming is completed on the upper pole surface

while the lower hemispherical shimming is completed on the lower pole

surface. Shimming of the equatorial plane is finished on both of upper

and lower poles.

According to Fig. 3, the values of Bm at the 13 points are measured

from top to bottom. Generally, the homogeneity of these 13 points is

very high, for example, the inhomogeneous value is only 7.699 ppm

based on expression (1). If all the magnetic field of the points on the

sphere is approached closely to the central value, the homogeneity of

the imaging region will meet the imaging requirement. Therefore, the

homogeneity of the central magnetic field provides a concrete reference

for establishing shimming model. The process of shimming based on

the central value is named as the central shimming strategy in this

paper.

The central shimming method based on integer programming is

proposed in this paper, which fulfills the combination of theory research

and practical manual shimming. The objective function of peak-topeak field tolerance subjected to series of inequalities is minimized,

while the numbers and locations of small shims are regarded as

objective variables. The possible positions of shims are previously set

and their final positions are determined by the number of shims on

corresponding position. When the shim number equals zero, no shim

is set on the position. The mathematic model is given as expression (3)

and (4).

MINIMIZE T

(3)

28

N

X

for j = 1, 2, . . . , M

Bij Xi + Bj Bm

i=1

N

st X

T

Bij Xi + Bj Bm

for j = 1, 2, . . . , M

i=1

ni Xi ni

for

(4)

i = 1, 2, . . . , N

the peak-to-peak magnetic field, whose expression is T = Bmax Bmin ;

Bmax and Bmin are the maximum and minimum of the magnetic flux

density respectively; Xi (being integer) is the number of the superposed

shims at location i; M and N are the total numbers of sampling and

shimming points respectively; Bj is the magnetic field value measured

at location j on the spherical surface of the imaging region before the

current shimming cycle, while Bm is the magnetic flux density at the

center of measured plane; Bij represents the change of the field at

point j caused by a shim of a fixed size placed at location i; and ni is

the most number of shims at a shimming location.

4. SIMULATION TEST OF SHIMMING

The magnetic field created by a shim with a certain size and set on a

certain position of a the pole surface is called as its contribution to

the original basic magnetic field, which is equal to the difference of the

magnetic field on the spherical surface before and after the placement

of the shim. In order to avoid the difficulty of analytic solution, all the

29

humidity and temperature.

The model described in Section 3 is solved using the branch and

bound method [79] based on the software LINGO8.0. The initial data

of magnetic field are measured in factory making magnets for MRI.

After gaining each shimming scheme, a simulation test is implemented.

The process of shimming simulation is divided into 3 steps. In order to

examine the results of every shimming step, numerical analysis method

is used to solve the whole contribution of the shim array.

Step 1: Shimming along 12 lines with an interval of 30 degree on

the pole surface (just shown as Fig. 5).

field with 15 interval.

Take the data of 156 sampling points firstly measured as the basic

magnetic field. Build the integer linear programming model according

to expression (3) and (4), and solve it to obtain the scheme of shimming

based on the software LINGO Fig. 4 shows the computation results,

which is used to direct the operator to do the shimming. Each small

ring represents a position where shims locate, and the figure around it

indicates the number of shims at the position in Fig. 4.

Step 2: Shimming along 12 lines with an interval of 30 degree and

beginning from the 15 degree on the pole surface (see Fig. 5).

As shown in Fig. 5, the dashed lines demote the new shimming

positions in this step. The new values of the basic magnetic field on the

spheric surface are obtained by the interpolation between the values

gained after Step 1. Then the integer programming model is solved to

obtain the 15-degree shims distribution on the radial beam.

Step 3: Shimming along 24 lines with an interval of 15 degree and

beginning from the 7.5 degree on the pole surface.

30

solving and shimming, the information of shims is obtained finally.

According to the final information of shims, which is obtained

based on the above steps, the practical shimming is operated through

sticking shims on both upper and lower poles. The shimming result is

obtained through measuring the magnetic field on the imaging sphere

surface with tesla meter.

In order to verify the availability of shimming algorithm in

advance, the contribution of shim array is computed using numerical

analysis. The difference between the magnetic field created by a small

shim and the main magnet is great. The former is only 107 T while

the latter is 0.3 T. Considering the rounding error and truncation error

of the finite element analysis, computation precision of the magnetic

field value will not be accurate if the permanent main magnet and iron

yoke are included in magnetic field analysis, so that the model of shim

array is built with the imaging region set as solving area.

Figure 6 shows the physical model of shimming scheme built with

the software ANSOFT [10].

Figure 7 and Fig. 8 are distributions of the magnetic flux density

of the basic field and the correction value after shimming for magnet 1.

Computation results of the shimming for two different magnets

of MRI are shown in Table 1 and Table 2 respectively. The

inhomogeneity is calculated by using expression (1). The 0 represents

the inhomogeneity of the basic magnetic field, and the 1 represents

the inhomogeneity after shimming. The represents the difference

of the inhomogeneity before and after shimming, whose expression is

= 0 1 . The improvment in the tables stands for the ratio of

to 0 .

31

Figure 7.

Distribution of the magnetic flux density and the

inhomogeneity of the basic field ( = 1813.0111 ppm).

Figure 8.

Distribution of the magnetic flux density and the

inhomogeneity after shimming ( = 860.16 ppm).

Table 1. Comparison of the inhomogeneity before and after shimming

(magnet 1).

Step

0 (ppm)

1 (ppm)

/0 (%)

Step 1

1813.01

1144.00

36.8

Step 2

1144.00

928

18.8

Step 3

928

860.16

7.31

32

(magnet 2).

Step

0 (ppm)

1 (ppm)

/0 (%)

Step 1

1081.82

911.14

15.78

Step 2

911.14

681.857

25.16

Step 3

681.857

637.46

6.5

Step

0 (ppm)

1 (ppm)

/0 (%)

Step 1

2002.75

1025.79

48.78

Step 2

1025.79

759.83

25.93

Step 3

759.83

463.62

38.98

The above tables indicate that the total homogeneity of two main

magnets after twice shimming is separately improved 52.56% and

41.08%. That is to say, the shimming schema of the central magnetic

field is effective. The information of shims can be obtained easily. The

homogeneity of different basic magnetic field can be improved in a wide

range.

In order to check the precision of the shimming algorithm proposed

in this paper, shimming experiment of the queen-post permanent

magnet is completed in the enterprise that produces the magnet used

for MRI. The results of shimming is shown in Table 3. Comparison is

done to the same magnet between the manual shimming method and

the central shimming method. The number of shims is 1400 for the

central shimming while the manual is nearly 10000; The time of the

central shimming is only 2 days, which includes time of computing just

3 minutes a time and setting the shims, while the manual shimming of

one proficient engineer costs more than 5 days.

5. CONCLUSION

In this paper, the central shimming method based on the magnetic

flux density of the sampling points, which are on the longitudinal

symmetrical axis in imaging region of MRI device, is proposed. The

method fulfills the combination of the integer linear programming and

the practical manual shimming process. The shimming scheme can

33

guide the manual shimming directly and efficiently. The validity of the

method is testified by numerical experiment and practical shimming

process. Corporation with the manufacturing enterprise shows that

the method proposed is quite efficient. The profound research of

the method proposed in this paper needs to be done to make the

homogeneity meet the imaging requirements

ACKNOWLEDGMENT

Supported by the Doctoral Program Foundation of Institutions of

Higher Education of China under Grant 20050142003.

REFERENCES

1. Wilkins, J. and S. Miller, The use of adaptive algorithms for

obtaining optimal electrical shimming in magnetic resonance

imaging (MRI), IEEE Trans. on Biomedical Engineering, Vol. 36,

No. 2, 202210, 1989.

2. Yamamoto, S., T. Yamada, M. Morita, et al., Field correction of

a high-homogeneous field superconducting magnet using a least

squares method, IEEE Trans. on Magnetics, Vol, 21, No. 2, 689

701, 1985.

3. Belov, A. and V. Bushuev, Passive shimming of the superconducting magnet for MRI, IEEE Trans. on Applied Superconductivity, Vol. 5, No. 2, 679681, 1995.

4. Lopez, H. S., F. Liu, E. Weber, et al., Passive shim design and

a shimming approach for biplanar permanent magnetic resonance

imaging magnets, IEEE Trans. on Magnetics, Vol. 44, No. 3,

394402, 2008.

5. Dorri, B. and M. E. Vermilyea, Passive shimming of mr magnets:

Algorithm, hardware, and results, IEEE Trans. on Applied

Superconductivity, Vol. 3, No. 1, 254257, 1993.

6. Wang, E., The passive shimming technique research for a special

application of the permanent magnet, Shenyang University of

Technology, Shenyang, China, 2007 (in Chinese).

7. Van de Panne, C. and F. Rahnamat, The first algorithm for linear

programming: An analysis of kantorovichs method, Ecnomics of

Planning, Vol. 19, No. 2, 7691, 1985.

8. Mitchell, J. E., Updating lower bounds when using karmarkars

projective algorithm for linear programming, Journal of

Optimization Theory and Application, Vol. 78, No. 1, 127142,

1993.

34

for a mixed integer branch-and-bound algorithm, Mathematical

Programming, No. 6, 212223, 1974.

10. Ansoft Corporation, Getting Started with Maxwell: Designing a

Rotational Actuator, Maxwell v11 Magnetostatic, 2005.

- Ultrasound 120925142801 Phpapp02Загружено:Swarn Avinash
- Alltexts.2011Загружено:Webster The-TechGuy Lungu
- Mri SyllabusЗагружено:brian2000
- EFW-2301 EFilm Workstation 4.2.x Release NotesЗагружено:Luis Adolfo Andonaegui Yanez
- Radiologi StrokeЗагружено:Ratna Ning Ayu Kustianti
- Anticancer Drug Testing- Proteome Signatures in Breast by MALDI, Immunostaining MRI-PET BiomarkerЗагружено:Rakesh Sharma
- msw_IЗагружено:Nikhil Kabbin
- Lecture18 Magnetic Magnetic Field of a Current Loop; Ampere's Law; Solenoids; ToroidsЗагружено:Taqi Shah
- Primary Hyperparathyroidism 1Загружено:Marco Meneses
- Brain Abscess Seminar 2000Загружено:VI Vqd
- dwillis form63Загружено:api-230026178
- Orden 2012 NeuroImageЗагружено:Michael
- Cristina 3Загружено:YuchungLee
- Kha ChanЗагружено:Sherif Adel
- WJR-6-794Загружено:dr arjun
- 1983-10_HP Journal PapersЗагружено:Elizabeth Williams
- Master Biomedical Engineering Guide 2015-2016 - Version 20151209Загружено:Antonis Nikolaidhs
- Radiology Procedure Manual-samplsЗагружено:dhiren2012
- MAGNET 2 .docxЗагружено:Sachin Yadav
- Lecture 1 Introduction Digital Image ProcessingЗагружено:Mona Sayed
- ajdr-08-02-30882Загружено:Alyaefkage
- TmjЗагружено:Ikrima Nsn
- combined answersЗагружено:api-359900292
- CT and MRI Improve Detection of Hepatocellular Carcinoma, Compared With Ultrasound Alone, In Patients With CirrhosisЗагружено:Fitri Yusnitasari
- Metastatic Bone Disease 2004Загружено:dianaveebae
- Imanes.pdfЗагружено:Pablo Torresi
- Head and Neck.pdfЗагружено:Rakesh Jadhav
- LED PhotoacousticЗагружено:Sang Trương
- Magnetic FieldЗагружено:It's Fritzzz
- Demir 2015Загружено:debby claudi

- The 1855 Weber-Kohlrausch Experiment (The Speed of Light)Загружено:Frederick David Tombe
- Ganguly Dynamics Complex NetworksЗагружено:Demian Pereira
- Aldrovandi R.-pereira J.G.-introduction to Geometrical Physics-World Scientific 1995Загружено:anthalia
- History of Medical ImagingЗагружено:Demian Pereira
- EMP GeneratorЗагружено:Red
- AUTOMATED BRAIN TUMUOR DETECTION AND IDENTIFICATION USING IMAGE PROCESSING AND PROBABILISTICS NEURAL NETWORK TECHNIQUESЗагружено:International Journal of Image Processing & Visual Communication (IJIPVC)
- Augmenting the Piezoelectric in Portland Cement Paste for StructuЗагружено:Demian Pereira
- Role of Dispersion of Multiwalled Carbon Nanotubes on Compressive Strength of Cement PasteЗагружено:Demian Pereira
- 2015_Rucka_Wilde_Ultrasound Monitoring for Evaluation of Damage in Reinforced ConcreteЗагружено:Demian Pereira
- NDT-SoilsЗагружено:Demian Pereira
- Concrete ReportЗагружено:Demian Pereira
- Defect Detection and Imaging From Phased ArrayЗагружено:Demian Pereira
- a445714-269Загружено:Demian Pereira
- 13a Reminiscent Study of Nature Inspired Computation Copyright IjaetЗагружено:Demian Pereira
- entropy-16-04497Загружено:Demian Pereira
- Multiparametric IschemiaЗагружено:Demian Pereira
- tgrowth-1Загружено:Demian Pereira
- Abdominal MRIЗагружено:Demian Pereira
- Difusion NeurosciencesЗагружено:Demian Pereira
- Lungs Qu Amous Cell CarcinomaЗагружено:Demian Pereira
- Glio Malig 1Загружено:Demian Pereira
- Diifusion BreastЗагружено:Demian Pereira
- Magnetic Exposure MeasurementЗагружено:Demian Pereira
- How Good is MRI-ProstateЗагружено:Demian Pereira
- Arxiv Lecture Notes in Quantum MechanicsЗагружено:Han Don
- Superluminal WavesЗагружено:Demian Pereira
- Aapm Basic DiagnosticЗагружено:Aisyah Muhaddisi
- A Model of Magnetic HyperthermiaЗагружено:Demian Pereira
- 10_De_GraafЗагружено:Demian Pereira

- Modern Mathematics in IndiaЗагружено:Adesh Naik
- Q121Загружено:hous
- Brain Mapping Final BrochureЗагружено:Isuru Chathuranga
- CONVERSION OF PLASTIC WASTE TO LIQUID FUELЗагружено:International Jpurnal Of Technical Research And Applications
- Intercultural Negotiation CasesЗагружено:Singh Ashu
- Prepking EE0-503 Exam QuestionsЗагружено:lemreno98
- Brushless Permanent Magnet Motor DrivesЗагружено:Yaswanth Kumar Lanka
- EntrepreneursЗагружено:Phyo Min Tun
- 1341903547012-E_chp8Загружено:Vijay Kumar
- Chin's Plot MethodЗагружено:Tin Aung Tun
- Transmisor de Temperatura Foxboro Rtt20Загружено:Felix Sepulveda
- ICQ is Your PCI DSS Compliance Program Working Correctly Res Eng 0216Загружено:junior_00
- Bba Syllabus II 2017-18Загружено:vini2710
- Kandhal Asphalt Literature for India Rev. 3 January 2019Загружено:Prof. Prithvi Singh Kandhal
- Review Paper IIЗагружено:Albyzia
- IHS Markit Standards Collections DocumentЗагружено:Erika Chong
- me203 hw1 fall13Загружено:Ammar Khan
- CATALOGO BucketelevatorsЗагружено:orlando_ubaldo
- Auditing in Erp EnvironmentЗагружено:dvbsca1722
- Material for Advance Ultra Supercritical Turbine Presentantion_20160420_1300A_FE0026294_EIOЗагружено:Mas Zuhad
- Unit Outline - EnS5257 Microprocessor SystemsЗагружено:Anup Vanarse
- Genral TyreЗагружено:sofian4sun
- Google SearchЗагружено:Lily
- SyllabusЗагружено:jalyssajax
- Air Brake ChamberЗагружено:alactraz
- QuietRock Acoustic Fire Design GuideЗагружено:JonEaton
- Betacam J Series ManualЗагружено:Steve Grajales
- IGI - Localization Algorithms (2009).pdfЗагружено:Cristina-Maria Dumitrascu
- aluЗагружено:amitcrathod
- 4450Загружено:Suneesh P